Granted Patent B2

US 12,548,057 · App. 17/562,823 · Granted Feb 10, 2026

Inventors: Chi-Chou Chiang (Hsinchu, TW); Hsin-Chien Huang (Hsinchu, TW)

Assignee: INDUSTRIAL TECHNOLOGY RESEARCH INSTITUTE

G06Q30/0631G06Q30/0201G06Q30/0255G06Q30/0641G06Q30/06313

Abstract

Claims (17)

1 . An assortment planning method for a smart store, comprising:

identifying, by an image recognition device, an infrared sensor, or a microwave sensor, a plurality of consumer tracks, wherein identifying the plurality of consumer tracks of a plurality of consumers includes:

step (1): detecting that one of the consumers enters the smart store;

step (2): taking a top view picture, identifying a human in the top view picture, and annotating the human in the top view picture;

step (3): taking other top view pictures every predetermined time, and continuously recording positions of the human in the top view pictures to obtain one of the consumer tracks;

step (4): returning to the steps (2) and (3) to obtain another one of the consumer tracks, when that another one of the consumers enters the smart store is detected;

step (5): recording leaving time of each of the consumers; and

step (6): recording ending time when all of the consumers leave;

detecting, by a weight sensor, the infrared sensor, a magnetron element, an image recognizer or an RFID sensor, a plurality of consumer interactive behaviors of a plurality of products, wherein the plurality of consumer interactive behaviors at least include a taking behavior and a putting back behavior;

associating the plurality of consumer interactive behaviors and the plurality of consumer tracks;

obtaining a plurality of consumption intentions including any instance of consumption intention or the plurality of consumption intentions of a plurality of products as referring to consumers' intent to purchase the plurality of products; and

determining, via a Multi-Objective Knapsack Problem (MOKP), a product combination of whole of the smart store according to a plurality of aggregate demands each of which includes a historical forecast demand, an out of stock demand and an alternative demands, wherein the out of stock demands and the alternative demands are obtained according to the plurality of consumption intentions, the product combination of whole of the smart store is determined by maximizing f(Σ j=1 n E(S j )*(p j −c j ), Σ j=1 n K j ) under a limitation that E j=1 n x j ≤number of product to list, E(S j ) is a forecast sale volume of each of the plurality of products, p j is a price of each of the plurality of products, c j is a cost of each of the plurality of products, K j is a performance indicator of each of the plurality of products, x j is a quantity of each the plurality of products, and (Σ j=1 n E(S j )*(p j −c j ) is a profit which is maximized in the MOKP.

2 . The assortment planning method for the smart store according to claim 1 , wherein the plurality of consumer interactive behaviors further include a viewing advertisement behavior, and a taking photo behavior.

3 . The assortment planning method for the smart store according to claim 1 , wherein the plurality of consumption intentions include a target product which is a product that one of the consumers is seeking, a substitute product, an interest product which is a product that is putted back by one of the consumers and is not been purchased, a sold out target product which is a product that is purchased by one of the consumers before but is sold out at that time.

4 . The assortment planning method for the smart store according to claim 1 , wherein the step of obtaining the plurality of consumption intentions includes:

screening the plurality of products to obtain at least one product corresponding to at least one of a viewing advertisement behavior, the putting back behavior, and a taking photo behavior; and

obtaining some of the plurality of products that were not purchased.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (40)